| Bug #32083 | server crashes on show status when InnoDB is not initialized | ||
|---|---|---|---|
| Submitted: | 3 Nov 2007 22:53 | Modified: | 25 Feb 2008 19:15 |
| Reporter: | Alexey Stroganov | Email Updates: | |
| Status: | Closed | Impact on me: | |
| Category: | MySQL Server: InnoDB storage engine | Severity: | S1 (Critical) |
| Version: | 5.0.50 | OS: | Any |
| Assigned to: | Timothy Smith | CPU Architecture: | Any |
[5 Nov 2007 15:13]
Heikki Tuuri
Aleksey, InnoDB really needs a writable tmpdir. It uses it as a buffer for some printfs. Assigning this to Marko. Maybe we could catch this and print an error message to the .err log? Regards, Heikki
[5 Nov 2007 16:21]
Alexey Stroganov
Heikki, Marko, probably we should somehow disable innodb engine in case of tmpdir is unwritable or at least assert but don't crash.
[7 Jan 2008 13:44]
Marko Mäkelä
This bug seems to have been introduced by someone at MySQL AB. The problem is that innodb_export_status is being called, even though InnoDB has not been initialized.
The fix is simple:
Index: ha_innodb.cc
===================================================================
--- ha_innodb.cc (revision 2222)
+++ ha_innodb.cc (working copy)
@@ -6339,7 +6339,9 @@ void
innodb_export_status(void)
/*======================*/
{
- srv_export_innodb_status();
+ if (innodb_inited) {
+ srv_export_innodb_status();
+ }
}
/****************************************************************************
[7 Jan 2008 13:58]
Marko Mäkelä
Sent innodb-5.0-ss2223.tar.gz to MySQL
[9 Jan 2008 17:41]
Timothy Smith
Workaround is to use a writable tmp_dir.
[24 Jan 2008 1:20]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/41186 ChangeSet@1.2588, 2008-01-23 18:20:04-07:00, tsmith@ramayana.hindu.god +1 -0 Applied innodb-5.0-ss2223 snapshot Fixes: Bug #32083: server crashes on show status when InnoDB is not initialized innodb_export_status(): Check that InnoDB has been initialized before invoking srv_export_innodb_status(). (Bug #32083) This bug does not exist in MySQL/InnoDB 5.1.
[24 Jan 2008 8:49]
Timothy Smith
patch queued to 5.0-build (doesn't affect other versions)
[25 Feb 2008 15:59]
Bugs System
Pushed into 5.1.24-rc
[25 Feb 2008 16:04]
Bugs System
Pushed into 5.0.58
[25 Feb 2008 16:05]
Bugs System
Pushed into 6.0.5-alpha
[25 Feb 2008 19:15]
Paul DuBois
Noted in 5.0.58 changelog. SHOW STATUS caused a server crash if InnoDB had not been initialized. Fix does not apply to 5.1+.
Description: I've started server and pointed tmpdir to location that is unwritable,executed 'show status' stmt and server crashes with following backtrace: (gdb) bt #0 0x00000000006b3a3a in srv_export_innodb_status () at srv0srv.c:1796 #1 0x000000000063d21f in innodb_export_status () at ha_innodb.cc:6310 #2 0x000000000062d04e in ha_update_statistics () at handler.cc:1159 #3 0x000000000065a9e2 in fill_status (thd=0x12cb6b0, tables=0x130a4e0, cond=0x0) at sql_show.cc:3573 #4 0x000000000065be45 in get_schema_tables_result (join=0x130ab20, executed_place=PROCESSED_BY_JOIN_EXEC) at sql_show.cc:4071 #5 0x00000000005b92df in JOIN::exec (this=0x130ab20) at sql_select.cc:1693 #6 0x00000000005baa8d in mysql_select (thd=0x12cb6b0, rref_pointer_array=0x12cd100, tables=0x130a4e0, wild_num=0, fields=@0x130ab20, conds=0x0, og_num=0, order=0x0, group=0x0, having=0x0, proc_param=0x0, select_options=2693024256, result=0x130ab00, unit=0x12ccb28, select_lex=0x12cced8) at sql_select.cc:2279 #7 0x00000000005b5e81 in handle_select (thd=0x12cb6b0, lex=0x12cca98, result=0x130ab00, setup_tables_done_option=0) at sql_select.cc:244 #8 0x0000000000582c16 in mysql_execute_command (thd=0x12cb6b0) at sql_parse.cc:2666 #9 0x0000000000588c3c in mysql_parse (thd=0x12cb6b0, inBuf=0x130a1c0 "show status like 'uptime'", length=25, found_semicolon=0x40881788) at sql_parse.cc:6094 #10 0x0000000000581331 in dispatch_command (command=COM_QUERY, thd=0x12cb6b0, packet=0x130a1d9 "", packet_length=26) at sql_parse.cc:1812 #11 0x0000000000580bbc in do_command (thd=0x12cb6b0) at sql_parse.cc:1586 #12 0x000000000057feb1 in handle_one_connection (arg=0xc24a10) at sql_parse.cc:1197 #13 0x00002b3f4ca14193 in start_thread () from /lib64/libpthread.so.0 #14 0x00002b3f4cf8945d in clone () from /lib64/libc.so.6 #15 0x0000000000000000 in ?? () mysqld.err file /data0/mysql-enterprise-5.0.50-linux-x86_64-glibc23/bin/mysqld-debug: Can't create/write to file '/root/tmp/ibACxcGs' (Errcode: 13) 071103 23:36:07 InnoDB: Error: unable to create temporary file; errno: 13 [New Thread 1082399040 (LWP 14567)] 071103 23:36:07 [Note] /data0/mysql-enterprise-5.0.50-linux-x86_64-glibc23/bin/mysqld-debug: ready for connections. Version: '5.0.50-enterprise-debug' socket: '/tmp/mysql.sock' port: 3306 MySQL Enterprise Server - Debug (Commercial) In the beginning of err file we see a warning that tmpdir is unwritable but unfortunately later in code this is not taken into account Also see BUG#15185 that should fix similar issue more than one year ago. How to repeat: start server with unwritable tmpdir: cd mysql-enterprise-5.0.50-linux-x86_64-glibc23 ./bin/mysqld --no-defaults --basedir=. --datadir=data/ --tmpdir=/root/tmp execute 'show status' stmt mysql -uroot -e'show status'